Korean hospitals are continuously investing in either their own information systems or a new innovative information technology for the purpose of improving health care quality and getting more cost effective ways for treating their patients. Currently, one of the biggest IT innovations by hospitals is electronic medical record (EMR). EMR usage percent was 20% in 2005, which was twice that of 1999 [1]. In this kind of innovation adoption, hospital management should evaluate how quickly their investments would achieve productivity and health care quality goals. An innovation adoption itself does not guarantee that organizations will achieve desired goals, unconditionally. For example, US business organizations had not found distinct evidence for productivity improvement until the 1990s, in spite of huge investments in information systems after 1970s, this is a phenomenon that has been called as productivity paradox [2].

It will be important to understand how IT innovations are adopted in either organization or individual levels, what kinds of mechanism are working in the adoption process, and why they want to either use or reject the innovations for their work purposes. Among these factors, employees' IT adoption or rejection behaviors should be the most important aspect, because the most urgent step for productivity enhancement starts from adopting and using any IT innovations. Moreover, in thinking about adoption behaviors, we need to identify what factors are the most important for adopting IT innovations.

The most eminent model on IT adoption behaviors is the technology acceptance model (TAM) [3,4], which is composed of two key determinants of behavioral intention, perceived usefulness and easy to use in terms of innovation characteristics. Using and applying TAM, many studies have investigated the determinants of behavioral intention for the IT of interest. In the Korean health care field, the studies [5-7] that were based on TAM reflected the key concepts of TAM relatively well and verified proposed path significances from perceived ease of use and perceived usefulness to usage intension. One study [8] adopted a variant approach, which assigned the final endogenous variable as satisfaction instead of behavioral intention. Moreover, rather than using behavioral intention as a final variable, one study [9] that investigated Web site usage behaviors tested and verified their hypotheses after specifying actual usage behaviors and loyalty to use as dependent variables and assigning usage intension as a mediator variable. However, even though all the above studies [3-9] considered two key innovation characteristics in their study models, they did not take into account individual innovativeness itself.

Rather than using Davis' key innovation characteristic variables, this study tires to testing the roles of individual innovativeness on IT adoption behaviors. Rogers' innovation diffusion theory [10,11] explains that personal traits as a predisposed individual characteristic is invariant across information technology types and is continuously influencing individual adoption decisions and behavior. Thus this study assumes that individual innovativeness directly determines actual IT adoption behaviors. In the most previous studies, usage intension was used as the dependent variable, because in behavioral intension could theoretically predict human behavior well. However, this study used actual adoption status as the dependant variable, which eliminates the assumption of the relationship between intension and behavior. This approach may give us more accurate results than using behavioral intension as the dependent variable. Moreover, without actual adoption and use of IT, the possibility for achieving an intended goal could not be expected.

For identifying the roles of individual innovativeness, this study selected e-purchasing behaviors and geographic positioning system (GPS) as the target technologies rather than the more current technologies with which subjects may not have adequate information. By studying the adoption of technologies that have been used for several years, the study subjects would be expected to have some information and experience with them. This allowed us to test the role of personal innovativeness on adoption behaviors in a natural social environment without intensive explanations or introduction sessions on the target technologies. Moreover, even this study selected two general IT innovations, the findings of this study may apply to understanding adoption and usage of physicians' EMR in hospitals because this study assumes that personal traits and character, personal innovativeness, can generally apply to adoption and usage behaviors of other IT innovations.

Based on Roger's [10,11] innovation diffusion theory, this study assumed that different innovative traits result in different IT adoption behaviors because each adopter group shares unique and common values, and differences of demographic characteristics also influence the differences of adoption behaviors and ACI. By selecting two different work professionals and IT types, this study also examined the invariance effects of ACI on adoption behaviors across two IT types. In the previous studies on IT adoption behaviors, demographic characteristics that influence IT usage were sex [13-15], age [13] and education level [16]. Even these previous studies were relatively more focused on explaining behavioral intensions rather than actual behaviors, this study adopted these variables as key components for explaining either actual usage behaviors or ACI. However, for explaining adoption status, all demographic characteristics of study I and II were not significant, but only age was significant in study I. In the merged data from study I and II, individual characteristics also were not significant and have no moderating effects with IT types. This result can be interpreted as their professional norms are more dominant over demographic differences in explaining individual innovativeness because legal requirements for getting their licenses and working within their professional job specifications may enhance their norms and make them more homogenous groups.

In explaining ACI, sex and marital status were significant. These results partially conform to the assertions of Roger [10], which insists that socioeconomic, communication behaviors and personality variables are related to ACI. Sex and marital status were significant, which can be linked to communication behaviors and personality. If males have more active communication networks and behaviors than females, males are more likely to be early majority than late majority compared to females, because Rogers [10] insists that communication behaviors are positively related to higher levels of ACI. Moreover, assuming the unmarried person are more likely to take a more favorable attitude toward change, science and education, Rogers' assertion was also verified with this result because of higher rates of innovative adopters in the unmarried persons, compared to the married.

This study identified that the differences of individual innovativeness determines IT adoption status after controlling demographic variables. In study I and II, innovative adopters were more likely to be adopters than early majority, and laggards were less likely to be current users than early majority. After considering IT types as a confounding variable, the significances of ACI were not changed. The odds of being current users in innovative adopters and early majority were higher than early majority and laggards, respectively. These results practically verify Roger's innovation diffusion theory [11] in Korean health care fields, which assumes that when people are exposed to new technologies or ideas, their underlying natures toward innovations thoroughly reflect their behaviors. Moreover, by identifying the similar results in study I and II, this study found the invariant effect of ACI on adoption behaviors.

In the previous studies, there was no study that examined the relationship between individual innovativeness and current adoption status. Yi et al. [17] identified the role of ACI for determining behavioral intensions. They found that ACI positively influences innovation characteristics (i.e., usefulness, ease of use and compatibility) that simultaneously determined behavioral intension, and ACI also directly influences behavioral intension, which implies that the effect of ACI was partially mediated through innovation characteristics. However, if we assume theoretically closer relevance between behavioral intension and actual behaviors, the results of Yi et al.'s study [17] and this study could be treated as a similar result. But this study attempt to eliminate the assumption by using actual adoption behaviors as the final endogenous variable. In a study using actual usage as the endogenous variable, Song et al. [9] used actual Web site usage. The Web site usage was determined by behavioral intension that was also influenced by usefulness. Thus this study, rather than testing innovation characteristics as a determinant of behavioral intensions, seeks to test how individual innovativeness itself works on adoption behaviors across IT types.

All moderating effects caused by IT types were not significant, which means that study subject technologies in each study are located at a similar position on IT adoption life cycle, and all demographic variables did not have any interaction effects with IT types. More specifically, the adoption rates of either each adopter category or demographic characteristics were not dramatically reversed by IT types. This result implies that ACI is the key determinant of adoption behaviors after considering the moderator (i.e., IT types) and individual demographic differences. In other words, regardless of IT types and demographic variables, ACI significantly influences personal IT adoption behaviors. These results partially conforms to the results of Yi et al.'s study [17], which tested moderating effects of ACI with main effects of innovation characteristics (i.e., usefulness, ease of use and compatibility), but this study sought to test the moderating effect of IT types with the main effect of ACI in explaining actual adoption behaviors. In both studies, there were no moderating effects.

For measuring individual innovativeness, this study used ACI, as Moore and Benbasat [18] proposed, which assumes that each individual has distinctive traits and belongs to one category in using IT innovations. In the other approach, Agarwal and Prasad [19] proposed five point Likert scales in terms of personal innovativeness in the domain of IT (PIIT). But this approach just gives the amounts of individual innovativeness, but not behavioral traits in a group. Thus, this study insists that ACI has more advantages for understanding IT adoption behaviors than PIIT. ACI can identify an individual's behavioral characteristics, which provide us with information about organizational IT implementation strategies for improving work efficiency.

The implication of these results is that individual innovativeness is a crucial element for understanding IT adoption behaviors. In a practical situation of EMR adoption and use, identifying individual innovative traits of health care professionals, hospitals may prepare and launch more sophisticated system diffusion strategies and system education classes, which can assure health care information system success. For example, if one organization, like the subjects of study I, has relatively more people in innovative adopters and early majority categories than the other categories, system education classes for their employees might be a more effective approach than a motivation program to use the new IT, because most employees are already internally motivated by themselves without extrinsic motivations in adopting and using the new IT. In the opposite situation, if there are significant portions of employees in late majority and laggards, the organization needs to design a motivation program for enhancing their adoption intensions and usages by introducing practical effects, gains, and instrumental functions after adopting and using the new IT.

In conclusion, this study verified, based on Roger's innovation diffusion theory, that adopter categories in innovations determined current adoption behaviors in the health care field. Moreover, this study also found that the influences of individual innovativeness exist in two types of information technology. In future studies, if more various types of IT are subject to testing, we would achieve more generalized results about the roles of individual innovativeness on IT adoption behaviors in health care fields. Moreover, a future study with Rogers' all personal innovative characteristics in study models, the findings might give us more broad perspectives on understandings IT innovations adoptions and usage behaviors.